Bipolar Ecl 1960s Deploy Ecl in Mainframe Computing

# Bipolar ECL Logic (1960s): Step 19 — Deploying ECL in Mainframe Computing

Where all of this circuit and system engineering actually went to work. This step surveys how IBM, Control Data Corporation (CDC), and later Cray Research used ECL in real mainframe and supercomputer designs through the 1960s and into the 1970s, as the concrete payoff for the speed-for-power-and-complexity tradeoff this series has worked through from the circuit level up.

## 1. IBM System/360 Model 91 and the High End of the 360 Line
IBM's System/360 Model 91 (1967), the fastest machine in the 360 family, used ECL-type current-switch circuitry (IBM's own in-house SLT and later ECL-derived circuit families) in its highest-speed logic paths — the arithmetic and instruction-pipeline circuitry where the nanosecond-class delay advantage over the saturating logic used in slower 360 models translated directly into instruction throughput. IBM's commitment to current-switch, non-saturating logic for its top-end mainframes through the late 1960s and 1970s mirrors, at the systems-house scale, the same tradeoff Motorola's MECL catalog parts offered commercially: accept higher power and tighter design discipline in exchange for clock rates saturating logic families could not reach.

## 2. Real Diagram: ECL Deployment Across 1960s-70s Mainframe and Supercomputer Lines

## 3. CDC and the Path Toward Cray
Control Data Corporation built its 6600 and 7600 systems (mid-to-late 1960s) around current-switch, ECL-style logic for the same reason IBM did at the high end of the 360 line — Seymour Cray's design philosophy throughout his CDC and later independent Cray Research career consistently prioritized raw clock speed over power efficiency, and ECL was the only commercially available logic technology through the 1970s that could deliver the clock rates his architectures targeted. This lineage runs directly into the liquid-cooled ECL systems referenced in Step 18 and continues past this series' 1960s window into the Cray-1 (1976) and subsequent Cray supercomputers, which remained committed to ECL logic until CMOS finally closed the speed gap in the 1990s — the subject of the closing step.

## 4. Companion Diagram: Stage Sequence and Trend Summary

## Deploying ECL in Mainframe Computing's Place in the Process Lineage
This is step 19 of 20 in the Bipolar ECL Logic (1960s) series, surveying how IBM, CDC, and the Cray lineage deployed ECL current-switch logic in the highest-performance subsystems of 1960s-70s mainframe and supercomputer design.